Equilibria shift toward lower volume states under pressure

The first result of pressurization on any system is compression; all atoms in the system move closer together, decreasing the system’s volume. Locally, however, some interatomic distances can increase, for example, when water molecules penetrate internal cavities and rearrange hydrogen-bond networks, so that the detailed structural response can be highly heterogeneous (Williamson et al., Reference Williamson2003). For systems in equilibrium between multiple states, pressurization will result as well in a shift in these equilibria towards the lower partial molar volume states, which for proteins are states that are less structured than the native state. The lower system volume of the less structured states of proteins stems from the fact that folded proteins are not perfectly packed (Frye and Royer, Reference Frye and Royer1998; Eriksen et al., Reference Eriksen2001; Williamson et al., Reference Williamson2003; Ando et al., Reference Ando2008; Roche et al., Reference Roche2012; Nucci et al., Reference Nucci2014; Maeno et al., Reference Maeno2015; Chen and Makhatadze, Reference Chen and Makhatadze2017). There exist within protein structures void spaces that are not occupied to any statistically significant degree by solvent (water) molecules. When these cavity-containing regions are disrupted, either the void volumes are eliminated and the volume taken up by solvent, or in some cases, the protein opens up and water fills the cavities. In both cases, void volume in the native state is occupied by water, lowering the partial molar volume of the system.

Although changes in internal cavities and packing defects appear to be the dominant contributors to the volume changes observed for protein unfolding and local structural transitions, other effects can also contribute. In particular, electrostriction of water around newly exposed charged or polar groups provides an additional negative contribution to the partial molar volume upon unfolding or local rearrangements (Murphy et al., Reference Murphy1998; Caro et al., Reference Caro2023). In most of the systems considered here, the magnitude and sequence-dependence of ΔV are best explained by the creation or elimination of internal voids; nonetheless, electrostrictive effects should be kept in mind, especially for conformational transitions that substantially alter solvent-exposed charge.

The differences in volume between native and fully unfolded states of proteins are relatively small, 1–2% of the total molecular volume of the protein (Royer, Reference Royer2002), necessitating significant pressures (1 bar-4 kbar is the typically accessible range) to elicit an effect. The pressure range over which protein unfolding occurs also depends upon the global stability, which can eventually be tuned by adjusting pH or temperature, adding denaturant, or even mutation in order to observe unfolding over the accessible pressure range. Volume changes for partial structural disruption or local conformational changes are even smaller than those for global unfolding, although the free energy differences are generally much smaller as well, such that intermediates are often detected. Moreover, locally less stable regions often are those that contain cavities and hence are more susceptible to pressure than better-packed regions. Note that, as discussed below, the volume change for protein unfolding (partial or global) is a strong function of temperature.

Activation volume (ΔV^#) as an analytic probe of protein conformational transitions

Analysis of the kinetics of conformational transitions provides information about the transition state of the reaction, and thus the mechanism for the transition. Pressure-jump relaxation experiments generally result in complete or partial disruption of protein structure. For simplicity, we will refer to these transitions as folding and unfolding. The pressure dependence of the observed rate constants can be analyzed to obtain the rate constants for folding and unfolding at atmospheric pressure, k°_F and k°_U, and the activation volumes for folding and unfolding, ΔV ^#_F and ΔV ^# _U. This allows one to place the partial molar volume of the transition state ensemble (TSE) relative to the folded and unfolded forms. We have referred to this as the Vi-value (Mitra et al., Reference Mitra2007), in analogy with the Φ-value for the activation m-values of denaturant-jump relaxation experiments (Fersht and Sato, Reference Fersht and Sato2004). Although Φ-values have provided many important insights into folding transition states, this approach also has limitations, as underscored by the Ψ-value analysis introduced by Sosnick and co-workers, which highlights cases where Φ-values can be misleading or ambiguous (Krantz and Sosnick, Reference Krantz and Sosnick2001). Moreover, the interpretation of Φ-values is predicated on the notion that the structure of the TSE remains the same, with only the stability changing upon mutation. Vi-value analysis offers a complementary volumetric perspective that can be particularly sensitive to cavity formation and hydration in the TSE. Indeed, Vi-value is a more structural as opposed to energetic parameter. We have found in several instances that the Vi-value and hence the volumetric (structural) properties of the TSE can be strongly modulated by mutation (Roche et al., Reference Roche2013).

In most cases studied to date (i.e., Vidugiris et al., Reference Vidugiris1995; Mohana-Borges et al., Reference Mohana-Borges1999; Pappenberger et al., Reference Pappenberger2000; Roumestand et al., Reference Roumestand2001; Seemann et al., Reference Seemann2001; Rouget et al., Reference Rouget2011), the partial molar volume of the TSE is situated between that of the folded and unfolded states, and generally closer to that of the folded state, as in the schematic derived from fluorescence-detected p-jump relaxation experiments on staphylococcal nuclease (Figure 1a). A relatively large positive activation volume for folding leads to significant slowing of folding under pressure, facilitating detection of the transition. A large, positive activation volume for folding also indicates that most of the cavities in the native state have formed at the barrier, corresponding to a relatively ‘late’ TSE. However, as noted for equilibrium volume changes associated with protein structural transitions (Seemann et al., Reference Seemann2001), the value of the activation volumes is strongly temperature dependent (Figure 1b) (Rouget et al., Reference Rouget2010). Thus, the placement of the volume of the TSE can change relative to the folded and unfolded states due to differential thermal expansion of the different states. In the case of the ankyrin repeat domain of the Notch protein, at higher temperature, the partial molar volume of the TSE was found to be higher than that of the folded state (Figure 1b), corresponding to a positive activation volume for unfolding as well. Likewise, small changes in sequence can modulate the absolute and relative volumetric properties of protein states.

Figure 1.

Volumetric diagram of protein folding reactions. (a) The volume diagram for the unfolding/folding reaction of staphylococcal nuclease (redrawn from Mitra et al., Reference Mitra2007) indicating the overall volume change between the folded and unfolded states ΔV _F (−75 ml/mol) and its decomposition into the activation volume for folding, ΔV ^#_F, (+65 ml/mol) and the activation volume for unfolding, ΔV ^#_U, (−10 ml/mol). (b) Temperature dependence of the volume diagram for the unfolding/folding reaction of the ankyrin repeat domain of the Notch receptor protein (redrawn from Rouget et al., Reference Rouget2011). Equilibrium volume changes, ΔV _F, are green, activation volumes for folding, ΔV ^#_F, are red, and activation volumes for unfolding, ΔV ^#_U, are blue. Note that the ΔV ^#_U is negative at low temperature and positive at high temperature due to the differences in thermal expansion of the folded state, α _F, with respect to the transition, α _T, and unfolded, α _U, states.

Intermediate timescale conformational exchange

While fast internal motions are for the most part largely pressure-insensitive, slower conformational transitions, on the microsecond to millisecond timescale, are exquisitely sensitive to pressure modulation, providing a window into protein substates and folding intermediates. Indeed, while the energetic barriers to these transitions are larger than those associated with ps-ns motions, the more extensive disruption involves larger differences in partial molar volume. Pressure-dependent ¹⁵N CPMG relaxation dispersion experiments have proven to be exceptionally powerful in dissecting the thermodynamics and kinetics of complex folding pathways. Bezsonova et al. (Reference Bezsonova2006) first demonstrated this approach using a millisecond-folding G48M mutant of the Fyn SH3 domain. By analyzing how relaxation dispersion profiles vary between 1 and 120 bar, they quantified the volumetric properties of both the unfolded and transition-state ensembles, revealing a substantial positive activation volume for folding (ΔV ^#_f) that reflects significantly decreased hydration in the transition state relative to the unfolded state. This pioneering work established the principle that pressure modulation of exchange rates in relaxation dispersion experiments affords precise insight into packing and solvation changes during folding transitions and underscores the utility of pressure perturbation to populate such excited states such that they can be characterized (Figure 3).

Figure 3.

High-pressure NMR methods for probing protein conformational exchange and folding kinetics. Schematic representation of how different high-pressure NMR approaches access distinct regions of the protein energy landscape. Local conformational exchange within the folded ensemble (F ↔ I) can be characterized using ¹⁵N CPMG relaxation dispersion and CEST experiments, which provide rate constants and populations of excited states. In contrast, real-time pressure-jump experiments – performed manually for slow transitions or using ultra-fast instrumentation for rapid events – along with ZZ-exchange spectroscopy, enable direct measurement of folding and unfolding kinetics (F ↔ U). Pressure shifts the free energy surface, stabilizing lower-volume states and altering the relative barriers between folded (F), intermediate (I), transition (T), and unfolded (U) states.

Building on this foundation, Tugarinov et al. (Reference Tugarinov2015) investigated a metastable mutant of Fyn SH3 with a three-state folding mechanism using HP relaxation dispersion. By globally fitting dispersion data across pressures, they extracted pressure-dependent rate constants, activation volumes for folding and unfolding, and changes in isothermal compressibility, enabling a fine reconstruction of the protein’s volumetric energy landscape. The results revealed that the on-pathway folding intermediate state, though energetically similar to the unfolded ensemble, has a partial volume close to the volume of the folded protein. Interestingly, the compressibility of the intermediate state was found to be larger than that of the folded state, indicative of an intermediate ensemble that is more structured than the unfolded ensemble but less rigid than the folded state.

More recently, Zhang et al. (Reference Zhang2022) applied pressure-dependent CPMG and CEST experiments to investigate conformational dynamics in the leucine-rich repeat protein pp32 (Figure 3). Mutational insertion of cavities at distinct locations revealed how pressure modulates exchange timescales and structural features of excited states. Specifically, a cavity introduced in the N-terminal capping motif (I7A) induced pressure-dependent exchange on the 500 μs-2 ms timescale detectable by CPMG experiments, while a cavity in the core region (L60A) led to slower exchange events (>2 ms) detectable by CEST. ¹⁵N chemical shift values obtained from the CEST experiments for the pressure-populated excited states revealed that the N-terminal region was unfolded, while the central core of the protein retained native-like structure. Very slow exchange for residues in the C-terminus of this variant of pp32 allowed mapping of the pressure-induced population of an intermediate with native-like structure in the C-terminal region by quantifying the corresponding peak intensities (Jenkins et al., Reference Jenkins2018). CEST was also useful in revealing the folded-like nature of the pressure-induced functional molten globule state of a small GTPase, Arf1 (Koduru et al., Reference Koduru2024). A note of caution: in the cases of the pp32 variants and the Arf proteins, the pressure-induced excited ensemble was very heterogeneous, with multiple transitions overlapping across the pressure range. This complexity precludes a full analysis of either the CPMG or CEST experiments to obtain exchange rates or from CPMG, excited state frequencies. In this case, it was useful to compare the T₂ relaxation time obtained from Hahn Echo experiments to that obtained via CPMG at a single, relatively high frequency, with differences being indicative of intermediate exchange.

As in the case of the Arf proteins, the enzymatic function of adenylate kinase involves a large conformational change. In the case of Adk this change corresponds to the opening of a lid obstructing the active site in the rate-limiting step. Comparison of the pressure effects on lid opening dynamics of Adk enzymes from mesophilic and piezophilic showed agreement between volumetric terms obtained from HP enzyme turnover and HP relaxation dispersion NMR (Stiller et al., Reference Stiller2019). Mutation of the mesophilic enzyme at three positions converted its pressure-independence to an equivalent negative activation volume for lid opening found for the piezophile, establishing the sequence determinants for the differential effects of pressure on dynamics. Generally, the studies discussed here illustrate how internal packing defects, when combined with pressure modulation, can unmask functionally relevant conformational exchange processes with distinct kinetics and spatial localization, reinforcing the sensitivity of relaxation dispersion to local energetic frustration and cooperative dynamics in repeat protein architectures. Although these excited states are preferentially populated under pressure, comparisons with temperature and denaturant perturbations, as well as with functional readouts, indicate that they correspond to low population substates that are already sampled at 1 bar and that modulate folding, allostery, and catalysis. Work from Akasaka, Kitahara and co-workers provides particularly clear examples of how pressure-stabilized states map onto the conformational landscape at 1 bar. In their PNAS study (Kitahara and Akasaka, Reference Kitahara and Akasaka2003), a pressure-stabilized intermediate was shown by HP NMR and thermodynamic analysis to correspond to the on-pathway folding intermediate inferred from kinetic experiments at ambient pressure, and subsequent work on lysozyme and ubiquitin generalized this picture to additional systems (Akasaka et al., Reference Akasaka2013). A series of HP NMR studies on ubiquitin further quantified the pressure-dependence of exchange rates and activation volumes between native-like substates, linking the amplitudes and timescales of backbone and side-chain fluctuations directly to functional motions (Kitahara et al., Reference Kitahara2005). Finally, in dihydrofolate reductase, pressure-dependent changes in NMR observables related low-population conformational excursions to specific steps in the catalytic cycle, demonstrating that the same fluctuations amplified by pressure underpin enzyme function at 1 bar (Kitahara et al., Reference Kitahara2000). Together, these studies make a strong case that HP NMR detects genuine folding intermediates and functional substates on the ambient-pressure landscape and provides quantitative access to their thermodynamic and kinetic parameters.

Pressure-dependent unfolding on a slow timescale monitored by NMR

When the pressure-induced unfolding of globular proteins occurs on a slow NMR timescale, the folding/unfolding equilibrium can be monitored by recording a series of ¹H-¹⁵N HSQC spectra at increasing pressure. Under such conditions, unfolding typically manifests as a progressive decrease in the intensity (or volume) of folded-state cross-peaks, and an increase in intensity for peaks of the unfolded state. Since locally, each residue backbone atom (NH or Cα) or side chain methyl group is either in its native environment or it is not, the residue-specific pressure dependence of these decreases in native state peak intensities can be fitted to a two-state transition to extract apparent residue-specific values of the folding free energy (ΔG _f) and volume change (ΔV _f). For an ideally cooperative two-state transition, cross-peak intensity curves for all residues should overlap and converge on a narrow range of ΔG _F and ΔV _F values within experimental uncertainty. Conversely, large variations in ΔG _F and/or ΔV _F across residues indicate deviations from cooperative behavior and suggest the presence of partially folded intermediates. Such analyses have revealed pressure-stabilized substates with reduced partial molar volume in multiple globular proteins, providing a sensitive probe of local folding cooperativity and packing defects (Roche and Royer, Reference Roche and Royer2018).

However, interpretation of intensity loss can be complicated when the unfolded state remains unassigned or invisible. In such cases, it is difficult to distinguish global unfolding from local unfolding events or from the transient population of structurally distinct high-energy substates that are broadened beyond detection. This challenge was exemplified in studies of the L99A mutant of T4 lysozyme. Wand and colleagues (Nucci et al., Reference Nucci2014) observed pressure-dependent disappearance of cross-peaks from the C-terminal domain, which they attributed to complete domain-scale unfolding. In contrast, Kitahara and coworkers (Maeno et al., Reference Maeno2015) proposed that these residues instead populated sparsely populated, pressure-stabilized high-energy states driven by the flipping-in of a surface aromatic side chain, reducing local cavity volume. By tracking both amide and methyl cross-peaks across pressure, they argued for the presence of a locally unfolded substate rather than a globally denatured form (Kitahara and Mulder, Reference Kitahara and Mulder2015). This interpretation was rebutted by Wand and Nucci (Reference Wand and Nucci2015) who cited SAXS and reverse micelle experiments to support the global unfolding hypothesis. Subsequent work by Xue et al. (Reference Xue2019) employed pressure-dependent CPMG relaxation dispersion and hydrogen exchange NMR to uncover multiple partly disordered excited states whose populations were tunable through internal cavity engineering.

An example of how HP NMR can be integrated with complementary structural techniques to resolve hidden folding intermediates is provided by Koduru et al. (Reference Koduru2024), who investigated the nucleotide-switch mechanism of the small GTPase Arf1. Using a combination of HP NMR, SAXS, FTIR, computation and site-directed mutagenesis, they identified a molten globule-like intermediate that primes Arf1 for GDP release and GTP binding. Pressure-induced changes in chemical shifts and peak intensities revealed a loosely packed, partially structured, and functionally relevant ensemble that remained undetectable at ambient pressure. Integration of the HP NMR data with SAXS, computation, and FTIR enabled a detailed view of the conformational landscape underlying functional switching. This study highlights the power of pressure perturbation, especially when combined with orthogonal techniques, to characterize low-lying intermediates involved in complex allosteric transitions and nucleotide-driven signaling cycles.

Real-time pressure-jump NMR

Beyond equilibrium measurements of folding intermediates and their interconversion dynamics, pressure perturbation can also be used to initiate and monitor time-resolved global un/folding reactions, offering a direct probe of folding kinetics and pathway selection. The large and positive activation volumes for folding (ΔV ^# _F) observed in many globular proteins can slow refolding by several orders of magnitude under elevated pressure. This kinetic deceleration offers a unique advantage for HP NMR spectroscopy: when the refolding timescale extends into the seconds-to-minutes range, real-time NMR measurements can be performed using conventional 2D experiments such as ¹H-¹⁵N HSQC or SOFAST-HMQC. Under such conditions, moderate pressure-jumps (typically 100–300 bar) enable direct, residue-specific monitoring of folding kinetics across the pressure-induced unfolding transition, permitting the dissection of folding pathways with near-atomic resolution (Figure 3).

Roche et al. (Reference Roche2013) applied this strategy to cavity-containing variants of Staphylococcal nuclease (SNase). Folding kinetics were tracked following pressure-jumps, using time-resolved series of 2D ¹H-¹⁵N spectra. Under optimized folding conditions, relaxation under pressure was slow enough (minutes to hours) that detailed residue-specific refolding profiles could be extracted from HP ¹H-¹⁵N HSQC or SOFAST HMQC spectra over time across multiple SNase variants and denaturant concentrations. This comprehensive dataset (>18,000 p-jump relaxation profiles) enabled precise mapping of the transition state ensemble (TSE) for the different variants, revealing how individual structural elements refold at distinct rates and how specific core mutations alter folding rates and routes. For instance, the Δ + PHS background variant, like WT SNase (Figure 1a) (Roche et al., Reference Roche2013), folded through a TSE featuring an ordered hydrophobic core and an unstructured C-terminal helix, while the Δ + PHS I92A mutant exhibited multiple parallel refolding pathways. Interestingly, the TSE for the folding of the I92A variant was very early, with only a small positive activation volume for folding compared to the WT (see Figure 1a). These results provided direct evidence that even subtle perturbations to packing can reshape the energy landscape and diversify folding mechanisms.

Analyzing the folding pathway of the immunoglobulin-like titin I27 domain, both as a single module and in tandem, Herrada et al. (Reference Herrada2018) took advantage of the time dependent increase of the methyl proton resonances corresponding to the unfolded state of the protein in a series of 1D spectra recorded after P-jumps, and found a very late, dehydrated, native like transition state. A subsequent study on another IgG domain, DEN4-ED3 from the Dengue virus, exploiting residue-specific unfolding rate constants determined using a similar approach (Figure 4), revealed that despite the structural homology to the titin I27 domain and the presence of a similar folding intermediate in the folding pathway, folding proceeded via a very early, wet TSE close in volume to the unfolded state (Saotome et al., Reference Saotome2019).

Figure 4.

Real-time 1H pressure-jump unfolding kinetics of the DEN4-ED3 domain from the Dengue virus protein envelope. (a) Time evolution of the unfolded-state methyl resonance intensity following a series of positive 200-bar pressure-jumps conducted between 700 and 2500 bar. The pressures indicated on the right correspond to the final pressures reached after each jump (e.g., “700 bar” represents a jump from 500 to 700 bar). (b) Chevron plot of the logarithm of the observed exchange times [ln(τ)] as a function of pressure, used to extract the global folding and unfolding rate constants (k _F and k _U), as well as the apparent activation volume of folding (ΔV ^#_F). Data adapted from Herrada et al. (Reference Herrada2018).

Together, these studies illustrate how real-time pressure-jump NMR leverages the naturally slow folding kinetics under high pressure that arise from positive activation volumes to provide mechanistic and structural insights into protein folding pathways. Crucially, this approach does not require specialized rapid-mixing equipment, making it broadly accessible for high-resolution dissection of folding landscapes in both simple and multidomain proteins.

Ultra-fast pressure-jump NMR

While conventional real-time pressure-jump NMR has proven effective for capturing slow folding kinetics, the structural interrogation of faster events, particularly those occurring on the millisecond to second timescale, requires much faster and precisely synchronized pressure changes. This technical challenge was first addressed by Kremer et al. (Reference Kremer2011), who developed a pioneering HP NMR system capable of delivering precisely controlled pulsed pressure perturbations within the NMR pulse sequence dead time. Their system, based on a microprocessor-controlled hydraulic pump, was integrated directly into the NMR pulse sequence and achieved jumps of ~80 MPa (800 bar) in under 30 milliseconds. Importantly, these pressure cycles could be applied repetitively and synchronized with spectroscopic acquisition, enabling kinetic studies and line shape analysis of pressure-perturbed protein transitions.

Bax and colleagues later introduced an improved system optimized for ultra-fast pressure drop experiments and applied it to monitor the folding of a pressure-sensitive ubiquitin variant (V17A/V26A) via hydrogen/deuterium exchange. In a landmark study, Alderson et al. (Reference Alderson2017) combined a rapid pressure release (from 2.5 kbar to 1 bar) with pulse-programmable acquisition to monitor hydrogen exchange protection within ~150 ms. By recording a series of 2D ¹H-¹⁵N spectra at defined delay times following the pressure drop, they tracked for individual backbone amides the switch of the exchange rates from unfolded to folded values, thereby mapping the sequence at which structural elements are formed during folding (Figure 3).

Charlier et al. (Reference Charlier2018) further refined the approach, adapting the pressure-jump protocol to monitor ¹⁵N chemical shift changes with high temporal resolution. This allowed them to follow the evolution of individual amide chemical environments immediately after pressure release via a pseudo-3D NMR experiment, enabling the identification of intermediate states transiently populated during the folding process. An interrupted folding protocol was subsequently introduced in a complementary study (Charlier et al., Reference Charlier2018) in which the protein was cycled between high- and low-pressure states within the duration of a single experiment. This approach successfully trapped a transient folding intermediate of ubiquitin that forms within the pressure cycle. The recovered ¹H, ¹⁵N, and ¹³C’ backbone chemical shifts revealed partial structuring of β-sheet elements and incomplete hydrogen-bond protection in terminal regions, consistent with the presence of a partially ordered folding intermediate that precedes full native-state acquisition. Importantly, this partially folded intermediate, only readily detectable under high pressure, is proposed to be an on-pathway species at atmospheric pressure, providing structural and mechanistic insight into the overall folding process. Furthermore, analysis of the kinetics suggested that the protein makes approximately 100 transitions to this intermediate state for every productive fold to the native state, highlighting the role of this high-energy state in kinetic partitioning.

In one recent study, Masoumzadeh et al. (Reference Masoumzadeh2024) applied the pressure-jump NMR approach to investigate proline isomerization as a rate-limiting step in folding of the L50A mutant of ubiquitin. By cycling between high- and low-pressure conditions, the authors could monitor proline isomerization kinetics in both model heptapeptides and the pressure-denatured protein. The data revealed site-specific differences in activation energies and pointed to metastable, partially folded states, notably a cis-E18-P19 isomer with unfolded C-terminal residues and exchange-broadened N-terminal signals, consistent with a transient, partially folded state.

Building on this body of work, Bax and colleagues solved the atomic-resolution structure of the transient, on-pathway folding intermediate of the pressure-sensitized ubiquitin mutant (Masoumzadeh et al., Reference Masoumzadeh2025). Using an extensive set of structural restraints, including chemical shifts, NOEs, and RDCs measured during synchronized pressure-jump cycles, the authors determined that the intermediate adopted a non-native β-sheet registry involving strands β1 and β5. This metastable conformation, which acts as a kinetic trap during folding, was previously implicated in the PINK1 mitophagy pathway, which underscores its functional relevance under ambient conditions. Analysis of long-timescale MD trajectories further revealed that such register-shifted β-sheet intermediates are visited repeatedly along the folding pathway, reinforcing the view that the pressure-populated intermediate corresponds to a genuine on-pathway state that is also sampled, at very low population, under native conditions.

Together, these studies established ultra-fast pressure-jump NMR as a cutting-edge method for accessing folding kinetics on the millisecond timescale, bridging the gap between traditional stopped-flow techniques and slower NMR-based pressure protocols. By combining pressure perturbation with advanced pulse programming and dedicated hardware, this approach offers real-time, residue-level snapshots of transient conformational states.

ZZ-Exchange NMR under pressure

When protein folding and unfolding transitions occur on the millisecond to second timescale and remain in the slow exchange regime on the NMR chemical shift timescale, two-dimensional ZZ-exchange spectroscopy (also called EXSY for exchange spectroscopy) offers a powerful route to directly measure forward and reverse folding rate constants. In an example of this approach (Zhang et al., Reference Zhang2016), Zhang et al. applied HP ZZ-exchange spectroscopy to the N-terminal domain of the ribosomal protein L9 (NTL9) that folds within 1 ms at atmospheric pressure and a destabilized variant of L9 C-terminal domain (CTL9 I98A) that folds within 70 ms. By tracking exchange between assigned folded and unfolded state resonances, residue-specific folding and unfolding rate constants were extracted as a function of pressure, allowing for the calculation of apparent residue-specific activation volumes for folding ΔV^#_F and unfolding ΔV^#_U. The results revealed that NTL9 undergoes folding through a near-ideal two-state mechanism with uniformly large ΔV^#_F and relatively small ΔV^#_U values. These findings are indicative of a compact, dehydrated transition state, with volumetric properties close to those of the folded state. While a similar pattern was observed for CTL9 I98A, several residues near the C-terminal β-strands exhibited faster-than-average folding rates, which points to transient intermediate states that may play a role in initiating folding. These findings underscore the utility of ZZ-exchange under pressure as a structurally sensitive probe of folding cooperativity and kinetic heterogeneity through direct observation of state-to-state transitions at equilibrium.